Dear Users
This week David Halliday from the Department of Electronics will give a
talk. David's talk will be followed by a YNiC project presentation by
Philip Quinlan and Nikos Andreadis titled "Switching tasks and
anticipating switching tasks: what is the brain doing?".
These events will take place at YNiC from 4-5 pm.
The title of David's talk is "The Noisy Brain – An introduction to time
and frequency domain signal processing techniques for neurophysiological
data".
http://www.elec.york.ac.uk/staff/dh20.html
Abstract: A common feature of neurophysiological signals is the presence
of noise, i.e. the signals have the appearance of random processes. In
addition, records taken under similar conditions do not have a similar
appearance. Analysis of such data therefore requires the use of some
form of statistical analysis. For example sample records of EEG recorded
from different scalp locations may not look alike, but a statistical
analysis may demonstrate that the records share a common underlying
rhythm. The field of time-series analysis provides a conceptual and
mathematical framework within which random signals, including
neurophysiological signals, can be analysed. Two approaches are
traditionally used in the analysis of time series – these are the time
and frequency domain approaches. The frequency domain approach is
generally based on Fourier methods. In the case of two or more
simultaneously recorded signals, an important concept is that of
correlation, the objective being to assess to what extent the activity
in one signal is correlated with that in a second signal. A key
frequency domain parameter in this context is the coherence function,
which provides a normative linear measure of association between two
time series, as a function of frequency. The talk will consider time and
frequency domain analyses of neurophysiological signals, including
coherence functions. These will be developed within a unified framework
where time and frequency domain techniques sit naturally alongside each
other, and can be used as complementary forms of analysis. A number of
extensions will be discussed, which look at alternative measures of
correlation (phase synchronization, causal measures); deal with time
dependency (Wavelets, Kalman filtering) and non linearities (higher
order spectra).
Everyone is welcome to attend.
Best wishes
Rebecca
--
************************************************************************
Dr. Rebecca E. Millman
Science Liaison Officer
York Neuroimaging Centre
The Biocentre
York Science Park
Heslington
YO10 5DG
Tel: +44 (0) 1904 567614
Fax: +44 (0) 1904 435356
